339f55ff65d244a8681e4879060274bfb6b9c1f1: Bug 1617183 - Remove cleanup code now that it's done. r=mdroettboom
Chris H-C <chutten@mozilla.com> - Fri, 28 Feb 2020 17:54:10 +0000 - rev 579143
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1617183 - Remove cleanup code now that it's done. r=mdroettboom This is a straight-up revert of bug 1611247 now that it's job is done. Differential Revision: https://phabricator.services.mozilla.com/D64780
0347e375abe34e2b4496d01bc007d22a6389cdc2: Bug 1616704 - Make `./mach build binaries && ./mach run` pick up binary changes to `libxul.so`. r=geckoview-reviewers,agi
Nick Alexander <nalexander@mozilla.com> - Wed, 26 Feb 2020 18:43:09 +0000 - rev 579142
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1616704 - Make `./mach build binaries && ./mach run` pick up binary changes to `libxul.so`. r=geckoview-reviewers,agi The Android-Gradle build plugin has evolved, as has GeckoView (it now expects libraries in `libs/` instead of `assets/`), so we need to ensure the libraries are copied into place at the right time. To test this, you can use an artifact build. Run `./mach build`, and then binary edit a string in `$TOPOBJDIR/dist/bin/libxul.so` (I used Emacs and `bbe` to do that). Run `./mach android build-geckoview_example`, and then verify the updated string has been packaged, for example with something like: ``` unzip -c $TOPOBJDIR/gradle/build/mobile/android/geckoview_example/outputs/apk/withGeckoBinaries/debug/geckoview_example-withGeckoBinaries-debug.apk lib/x86_64/libxul.so | strings | grep NEW 'STRING' ``` Coincidentally, this is essentially the same issue as https://github.com/mozilla/application-services/issues/2659. Differential Revision: https://phabricator.services.mozilla.com/D64430
154826a580cfa74f9e07521a1c696d08d9f66cdc: Bug 1615001 - `always-target` android-lints like other mozlints. r=tomprince
Nick Alexander <nalexander@mozilla.com> - Fri, 28 Feb 2020 18:34:19 +0000 - rev 579141
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1615001 - `always-target` android-lints like other mozlints. r=tomprince Differential Revision: https://phabricator.services.mozilla.com/D64784
b7ab4c947688fab1ee8de2ef13a6033e2307d590: Bug 1596513: Part 3: Ensure drop shadow blur radius does not exceed MAX_BLUR after scale factors are applied r=gfx-reviewers,nical
cbrewster <connorbrewster@yahoo.com> - Fri, 28 Feb 2020 17:39:05 +0000 - rev 579140
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1596513: Part 3: Ensure drop shadow blur radius does not exceed MAX_BLUR after scale factors are applied r=gfx-reviewers,nical This was causing one of the large drop-shadow wrench reftests to timeout. This is only a partial fix, as we should be checking the scale factors earlier on when sanitizing the filter input. This will ensure we match what the non-WR backend is doing and will prevent overinflation. Differential Revision: https://phabricator.services.mozilla.com/D64197
ce1593bf9d51fa70cf7823c2aea5a760ade1df81: Bug 1596513: Part 2: Allow for floating point dx/dy on feDropShadow attributes r=nical
cbrewster <connorbrewster@yahoo.com> - Fri, 28 Feb 2020 17:10:17 +0000 - rev 579139
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1596513: Part 2: Allow for floating point dx/dy on feDropShadow attributes r=nical WebRender expects drop shadow dx/dy to be in user space, which is a floating point value. Before this, a dx/dy such as (0.2, 0.2) was truncated to (0, 0) resulting in no offset of the shadow. Differential Revision: https://phabricator.services.mozilla.com/D63442
a9d8b62a399349559ceff12111f5d7fd3fab4acc: Bug 1596513: Part 1: Take scale factors into account when rendering drop shadows in WebRender r=gfx-reviewers,kvark
cbrewster <connorbrewster@yahoo.com> - Fri, 28 Feb 2020 17:09:44 +0000 - rev 579138
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1596513: Part 1: Take scale factors into account when rendering drop shadows in WebRender r=gfx-reviewers,kvark Differential Revision: https://phabricator.services.mozilla.com/D63441
f3fdc0edd2c4853f93fbc549f69823c657ef0e67: Bug 1618526: Don't check for duplicate parameters when duplicates aren't allowed in the function context. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 17:26:48 +0000 - rev 579137
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1618526: Don't check for duplicate parameters when duplicates aren't allowed in the function context. r=jandem Differential Revision: https://phabricator.services.mozilla.com/D64581
098eb676052f87dbc5da59aaa8ee1cd99b26b7da: Bug 1380281 - Part 3: Only create an unmapped arguments object when the arguments contents are observed. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 17:26:40 +0000 - rev 579136
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1380281 - Part 3: Only create an unmapped arguments object when the arguments contents are observed. r=jandem When the script contains `JSOp::SetArg` and uses unmapped arguments objects, we can still omit creating the arguments object as long as the arguments contents are never observed. Differential Revision: https://phabricator.services.mozilla.com/D64576
6a01bbfc11f77be6258af76ed6af5405df78995e: Bug 1380281 - Part 2: Decide within the arguments analysis if unmapped arguments require an arguments object. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 15:00:25 +0000 - rev 579135
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1380281 - Part 2: Decide within the arguments analysis if unmapped arguments require an arguments object. r=jandem This change just moves the point where we decide if parameter modification combined with an unmapped arguments object requires to eagerly create an arguments object. It shouldn't result in any change of behaviour. Differential Revision: https://phabricator.services.mozilla.com/D64575
ddd4454e71c09b697ddff610944981c99cdbe2f4: Bug 1380281 - Part 1: Support JSOp::SetArg with unmapped arguments object. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 15:31:58 +0000 - rev 579134
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1380281 - Part 1: Support JSOp::SetArg with unmapped arguments object. r=jandem When the arguments object is unmapped, JSOp::SetArg can directly write to the arguments slot, because we don't need to synchronise it with the arguments object. Differential Revision: https://phabricator.services.mozilla.com/D64574
c367dad821298ba7098c4a00bf3a5593ec61e53c: Bug 1478350 - Part 7: Update a method name in a comment. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 14:59:31 +0000 - rev 579133
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1478350 - Part 7: Update a method name in a comment. r=jandem Looks like this should be `unaliasedActual()`, when interpreting this changeset correctly: https://hg.mozilla.org/mozilla-central/rev/adcd5d3c984efef3688189b11c6ec6276ad014c8 Drive-by change: Remove a stray, commented out code line. Differential Revision: https://phabricator.services.mozilla.com/D64573
699e15941bae4650ddba8023ca1e67a5d60c30b4: Bug 1478350 - Part 6: Remove AliasSet::FrameArgument. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 14:59:05 +0000 - rev 579132
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1478350 - Part 6: Remove AliasSet::FrameArgument. r=jandem We don't create lazy arguments when `JSOp::SetArg` is present, which means `modifiesArguments()` is always `false` in this method. Differential Revision: https://phabricator.services.mozilla.com/D64572
84d19c5c6a7aeaf2bfccc4faddd89a928f951cf3: Bug 1478350 - Part 5: Remove an outdated comment in IonBuilder::jsop_setarg. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 14:58:40 +0000 - rev 579131
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1478350 - Part 5: Remove an outdated comment in IonBuilder::jsop_setarg. r=jandem This comment dates back to before bug 921120 and bug 995336 were written and no longer makes any sense with the current implementation. Differential Revision: https://phabricator.services.mozilla.com/D64571
b89721d62b74671a11a3562e2eac2a0f3ea1b033: Bug 1478350 - Part 4: Remove no longer used MSetFrameArgument node. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 14:58:12 +0000 - rev 579130
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1478350 - Part 4: Remove no longer used MSetFrameArgument node. r=jandem Part 3 removed the only place where MSetFrameArgument instructions were created. Differential Revision: https://phabricator.services.mozilla.com/D64570
8a3508d85549ea256dc2bec8f8d41b0ee151b2fe: Bug 1478350 - Part 3: Restructure IonBuilder::jsop_setarg. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 14:57:38 +0000 - rev 579129
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1478350 - Part 3: Restructure IonBuilder::jsop_setarg. r=jandem Restructure `jsop_setarg()` to make it clearer when we abort Ion compilation. The `MSetFrameArgument` code added in bug 921120 (and later disabled in bug 957475), was written before bug 995336, so it did handle a case which can no longer happen, namely `JSOp::SetArg` bytecode operations on lazy arguments. If we want to support the use case from bug 921120, we will need to update the arguments analysis first. The restructuring happened as follows: `info().argsObjAliasesFormals()` and `info().argumentsAliasesFormals()` both imply `info().hasArguments()`, so when we handle `!info().hasArguments()` first, we just need to emit `current->setArg(arg)` and can then directly return. Furthermore `info().argsObjAliasesFormals()` and `info().argumentsAliasesFormals()` also imply mapped arguments semantics, so we handle `!info().hasMappedArgsObj()` next and abort the Ion compilation. This is just done to match the current behaviour and should be removed in a follow-up patch! When we're running the arguments analysis, `info().needsArgsObj()` is true, so we always end up with emitting `MSetArgumentsObjectArg`. The arguments analysis will then flag the arguments object as being needed, cf. `ArgumentsUseCanBeLazy()`. That means `info().needsArgsObj()` is always true at this point, whether or not we're currently running the arguments analysis. Differential Revision: https://phabricator.services.mozilla.com/D64569
3d52ff211475831a2b0484e6f9273209066ceec6: Bug 1478350 - Part 2: Simplify return statement in IonBuilder::jsop_setarg. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 14:57:15 +0000 - rev 579128
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1478350 - Part 2: Simplify return statement in IonBuilder::jsop_setarg. r=jandem Drive-by change: `GET_ARGNO(pc)` is passed to the function as `arg`, so we don't need to recompute it again within the function. Differential Revision: https://phabricator.services.mozilla.com/D64568
b0b1bbf5387c69dddc96d32443812e9d5c7a98f2: Bug 1478350 - Part 1: Move JSOp::GetArg handling into a separate function. r=jandem
André Bargull <andre.bargull@gmail.com> - Fri, 28 Feb 2020 14:56:42 +0000 - rev 579127
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1478350 - Part 1: Move JSOp::GetArg handling into a separate function. r=jandem Add IonBuilder::jsop_getarg next to the existing IonBuilder::jsop_setarg, so we have both operations next to each other, which makes it easier to compare how aliased arguments are handled. Differential Revision: https://phabricator.services.mozilla.com/D64567
991ef93c0856c10d19d49052f705c3af861e125d: Bug 1618903 - Stack grid item rule shouldn't apply to anonymous children. r=dholbert
Emilio Cobos Álvarez <emilio@crisal.io> - Fri, 28 Feb 2020 18:33:56 +0000 - rev 579126
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1618903 - Stack grid item rule shouldn't apply to anonymous children. r=dholbert Differential Revision: https://phabricator.services.mozilla.com/D64786
7c6bdadeea29fab3700320515798d84ba3414f4f: Bug 1613348 - Turn on the modern Search Service configuration for nightly builds. r=daleharvey
Mark Banner <standard8@mozilla.com> - Fri, 28 Feb 2020 12:13:50 +0000 - rev 579125
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1613348 - Turn on the modern Search Service configuration for nightly builds. r=daleharvey Differential Revision: https://phabricator.services.mozilla.com/D63162
4e986d118c340585b5f4c7df3481c03573c5e60c: Bug 1618935 - Add named threads used by Thunderbird to ThreadAllows.txt. r=KrisWright
Rob Lemley <rob@thunderbird.net> - Fri, 28 Feb 2020 18:23:15 +0000 - rev 579124
Push 12808 by ffxbld-merge at Mon, 09 Mar 2020 15:29:44 +0000
Bug 1618935 - Add named threads used by Thunderbird to ThreadAllows.txt. r=KrisWright This is to fix build bustage until Thunderbird engineers are able to migrate these threads to use background tasks. Differential Revision: https://phabricator.services.mozilla.com/D64795
(0) -300000 -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 tip